google-videos-search

Pass

Audited by Gen Agent Trust Hub on Jul 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill instructions and supporting scripts were analyzed and no evidence of malicious behavior, obfuscation, or unauthorized data exfiltration was found. The skill operates within its stated purpose of providing a wrapper for the Frevana Google Videos service.
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by ingesting content from an external search API, which is a standard risk for integration skills.
  • Ingestion points: Search results are fetched from the Frevana API at ai-factory.frevana.com within scripts/search_google_videos.sh.
  • Boundary markers: None are defined in the instructions to separate API content from agent instructions.
  • Capability inventory: The scripts/search_google_videos.sh script possesses the capability to write output to a user-defined local file path and print data to standard output.
  • Sanitization: The script performs structural validation to ensure the response is valid JSON before it is passed to the agent context.
  • [COMMAND_EXECUTION]: The bash script uses inline Python execution blocks for processing command-line arguments and validating JSON responses. This logic is local, follows best practices for input handling, and does not incorporate unsanitized remote data into the execution flow.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 13, 2026, 09:15 PM
Security Audit — agent-trust-hub — google-videos-search